store: Restic backup stale lock fix and Discord alerting

graph/fixes/restic-backup-stale-lock-fix-and-discord-alerting-9f55a3.md

@@ -0,0 +1,29 @@
+---
+id: 9f55a3cb-3af3-4389-b255-04e90b396988
+type: fix
+title: "Restic backup stale lock fix and Discord alerting"
+tags: [homelab, restic, backup, discord]
+importance: 0.8
+confidence: 0.8
+created: "2026-02-19T13:22:09.621843+00:00"
+updated: "2026-02-19T13:22:09.621843+00:00"
+---
+
+## Problem
+Restic backup on nobara-desktop had a stale lock from 2025-12-09 that prevented forget --prune from running for 2+ months. Result: 73 snapshots accumulated instead of ~18. Backup data was fine (dedup kept disk at 219 GiB) but retention never ran.
+
+## Root Cause
+The backup script used set -euo pipefail but had no lock cleanup or failure alerting. When restic forget hit the stale lock, it exited with code 11 silently every night.
+
+## Fix Applied (2026-02-19)
+1. Cleared stale lock: restic unlock --remove-all
+2. Updated /home/cal/.local/bin/restic-backup.sh with:
+   - Pre-backup restic unlock --remove-all to clear stale locks automatically
+   - Discord webhook alerting (Homelab Alerts channel) on any failure
+   - trap ERR to catch unexpected exits
+   - --cleanup-cache flag on backup command
+3. Created backups/CONTEXT.md with full documentation
+4. Updated CLAUDE.md loading table with backup keywords
+
+## Key Lesson
+Any automated task with set -e needs failure alerting. Silent failures can go unnoticed for months. Always pair scheduled tasks with notification on failure.